Applications of Thermal Mechanical Compression Tests in Food Powder Analysis

A new Thermal Mechanical Compression Test (TMCT) was applied for glass-rubber transition and melting analyses of food powders and crystals. The TMCT technique measures the phase change of a material based on mechanical changes during the transition. Whey, honey, and apple juice powders were analyzed for their glass-rubber transition temperatures. Sucrose and glucose monohydrate crystals were analyzed for their melting temperatures. The results were compared to the values obtained by conventional DSC and TMA techniques. The new TMCT technique provided the results that were very close to the conventional techniques. This technique can be an alternative to analyze glass-rubber transition of food, pharmaceutical, and chemical dry products.

Keywords:

• Stickiness
• Glass transition
• Mechanical test
• Melting
• Food powders
• Sucrose
• Glucose monohydrate

INTRODUCTION

The glass transition can be measured based on a change in heat capacity using a thermal analysis instrument such as Differential Scanning Calorimetry (DSC) or by a change in mechanical properties using a thermal mechanical analysis instrument such as Dynamic Mechanical Thermal Analysis (DMTA) or Thermal Mechanical Analysis (TMA). The mechanical tests are normally more sensitive for measurement of the glass transition than the thermal analysis, particularly for the food polymers (e.g. proteins, starches) whose change in heat capacity is very small during this transition but their mechanical property change is very large, thus easy to detect.[1–3]

The conventional DMTA or TMA mechanical testing methods are more suitable for solid or semisolid materials than powders. There has not been a report on DMTA direct applications for food materials in powder form. It is common practice that the powdered materials are moulded or compressed into certain solid shapes such as a tablet, strip, or sheet to facilitate the grasp the sample in place during compression and tension.[4–12] An extra sample preparation step is required to be compressed into a sheet or tablet form, and it may generate changes of the sample properties, as compared to free-flowing bulk samples. In addition, these techniques are found to be unsuitable for applications in food analysis due to possible moisture losses during the test. The moisture content of the sample being analyzed greatly affects the mechanical properties, especially when the temperature increases to very high level (i.e. beyond T_g).[10] This limits DMTA applications in food analysis where moisture is a significant component. Evaporation loss may be avoided by using high pressure cell, however, this is not an easy solution. A lag between the measured temperature and the observed effects would occur because heating is controlled by the test chamber temperature not the actual sample temperature. Therefore, the determined T_g can be much higher than the actual values. A slow heating rate (i.e., 2°C/min) could be considered to obtain accurate T_g values, however, it may cause moisture variation in the sample, and it is time consuming—which is not suitable for routine applications where instant results are required for operation control. The applying force is also very small (i.e., limited to 15 N maximum for TMA test), which limits the sensitivity and applicability of the technique. These techniques are also very expensive and require the use of costly consumables for the analysis of powder samples.

Because of the previously mentioned limitations of sophisticated conventional techniques, a simple Thermal Mechanical Compression Test (TMCT) technique was developed for the analysis of food powders.[5] This new instrument measures glass-rubber transition temperature (T_g-r) based on changes in mechanical behavior during thermal compression tests in creep mode. It has been tested and validated against standard DSC and DMTA/TMA techniques. The design and its validation to measure the T_g-r of skim milk powder are described elsewhere.[1] The schematic diagram of the testing temperature controlled block is depicted in Fig. 1. It was also found that this technique was simple, cheap, user friendly, accurate and fast. It is, therefore, very important that this technique be explored further to determine its potential applications that will be of great benefit to food, pharmaceutical, and chemical industries. The objective of this article was to investigate the applicability of this technique to other types of food powders; in particular, potentially sticky materials such as whey, honey, and apple juice powders, as well as some sugar crystals such as glucose and sucrose.

Figure 1. Schematic diagram of design of the TMCT testing device, which is attached to TA.XT2 texture analyzer.

METHODOLOGY

Sample Preparation and Spray Drying

Whey, honey, and apple juice were spray dried and subsequently analyzed in this work. Three 1.5 kg batches of 30% w/w whey solution were prepared from commercial whey powder (Dairy Farmers Co., NSW, Australia) using 50°C distilled water. Major compositions of the whey powder provided by the supplier were < 3 kg/100 kg moisture, 2.5 kg/100 kg protein and 75 kg/100 kg lactose. The solutions were maintained at 50°C until spray drying, which was carried out using inlet and outlet air temperatures of 180 ± 1°C and 80 ± 2°C, respectively. Because honey is mainly composed of fructose and glucose, with their characteristic low T_g, it was not possible to spray dry without a stickiness problem. Recommendations have been made on the use of high T_g additives to help increase the product T_g and hence reduce stickiness.[6] Maltodextrin DE6 (Roquette Freres, France) was incorporated as 50% of the total solids with honey (82% solids) in a 45% total solids solution. Honey used in this study was Australian eucalyptus honey named Napunyah. The solution was heated to 45–50°C for its preparation, and it was maintained at this temperature through out the drying process. The solution was spray dried at 150 ± 1°C and 65 ± 2°C inlet and outlet air temperatures, respectively.

Similar to the preparation of the honey solution, the maltodextrin DE6 was added to the apple juice solution at 50:50 (maltodextrin:apple juice solids) ratio. The original solids content in concentrated apple juice (Golden Circle, Brisbane, Australia) was 70.5–71.5% w/w with 2.3–2.8% acidity (expressed as anhydrous citric acid). Distilled water was used for the dilution. A batch of 1.5 kg of 45% w/w solution was made and spray dried at 150 ± 1°C and 65 ± 2°C inlet and outlet air temperatures, respectively; the temperature of the feed solution was maintained at 45–50°C.

Spray drying of all samples was carried out using Anhydro Lab 1 spray dryer (Copenhagen, Denmark) with a rotary atomizer. Spray drying was carried out using three batches of 1.5 kg solution for each sample. The drying air flow rate and speed of the atomizer were set at a constant for all samples at 0.05 m^3/s and 20,000 rpm, respectively. The sample was collected from the cyclone separator into a pre-weighed glass bottle. This was the sample used for further analysis. At the end of the drying cycle, the deposited powder in the drying chamber was recovered into another bottle.

Moisture Content and Water Activity Analyses

Moisture contents of whey powder were analyzed according to the standard AOAC method.[14] The vacuum oven method was also used for the determination of moisture contents in honey and apple juice powders, but a lower temperature (70°C as compared to 100°C for whey powder) was used to avoid burning of the sugar components, which may generate errors. Percentage moisture content was calculated on a dry basis based on the amount of water evaporated. Water activity of the samples was measured using an AquaLab 3 water activity meter (Decagon Devices, Inc., Pullman, USA). Separate moisture and water activity measurements were conducted for each batch. A sorption isotherm of whey powder was carried out at 20°C using an evacuated method.[15–16] This was done to obtain the powder at different moisture contents for TMCT measurement.

Glass Transition Analysis Using TMCT and Standard Conventional Methods

The glass-rubber transition temperatures (T_g-r) of powder samples were analyzed using the TMCT technique. For each sample, 1 g of powder was tested using a 29.43 N (3 kg) compression force, and a 30°C/min heating rate. The heating rate was digitally controlled by a BTC9300 heater controller (Brainchild Electronic Co., Taipei, Taiwan), which was connected to the sample cell. These testing parameters were selected within the recommended ranges.[1] The standard material used for correction was maltodextrin powder. The influence of moisture content on T_g-r was also determined using samples equilibrated to different moisture contents. The analysis was done in triplicate.

The standard DSC and TMA techniques were used for comparison purposes. For DSC analysis (Perkin Elmer, U.S.), about 5–10 mg of the powder sample was weighed into a 50 μl DSC aluminium pan and press sealed with a lid using a DSC sample press. Thermal scanning of the sample was carried out using, a sealed empty pan as a reference, in 4 steps: 1) Isothermal at −20°C for 1 minute; 2) Heating from −20 to 90°C at 10°C/min; 3) Cooling from 90 to −20°C at 50°C/min; and 4) Heating from −20 to 120°C at 10°C/min. T_g was determined for an onset, midpoint, and endpoint from the second heating thermogram. At least one duplicate analysis was completed for each sample.

Analysis of T_onset was also conducted using TMA (DMTA MK IV, Rheometric Scientific, Piscataway, NJ) technique at a heating rate of 5°C/min, and scanning from 25–120°C.[1] A 25 μl DSC pan was filled with approximately 15 mg of powder sample, and a lid was gently pressed to slide it down into the pan. A static creep test was performed using a 5 mm compression probe and 9.81 N compression force. The sample under compression was maintained at 25°C for 300 s before heat scanning. The T_onset was analyzed using the DMTA software. The analysis was carried out with a minimum of two replicates for each sample.

Melting Point Test of Sucrose and Glucose Monohydrate Crystals

Beside amorphous powders, crystalline products also contribute a considerably large volume of food commodities. Knowing their melting temperatures would, therefore, facilitate process designs and operation. Therefore, common food crystals, such as sucrose and glucose, were selected to demonstrate the capability of the TMCT for melting point analysis. Because melting of crystals exhibits a sharp change in mechanical properties,[1,17] it is therefore possible to use as little amount as 0.5 g to as much as 2 g of sample for the test without significant interference of the sample cell expansion. However, a sample size of 1 g was used for consistent results. The analysis was carried out using a 29.43 N compression force and 30°C/min heating rate. Triplicate measurements were conducted for each sample. DSC analysis was also performed using a standard heating rate of 10°C/min to determine the melting points of the crystals and for comparison with the results obtained from the TMCT technique. It was expected that the mechanical change during the melting of these crystal would occur only once as they are single (pure) solute.[18]

RESULTS AND DISCUSSION

Moisture Content and Glass-Rubber Transition Temperatures of Food Powders

The glass-rubber transition temperatures (T_g-r) for each powder samples were determined using the TMCT technique. Fig. 2 shows a typical mechanical response as a function of temperature. The glass-rubber transition is determined using a linear regression method in the range of increasing in compression probe displacement. The samples of whey powder without lactose crystallization and equilibrated at 20°C were analyzed, and the results are presented in Table 1, as a function of moisture content. For honey and apple juice powder, only the original powder obtained from spray drying was analyzed; the results are shown in Table 2. It can be seen that the TMCT technique was very precise, shown by very low standard deviations.

Figure 2. A typical TMCT thermal mechanical response during glass-rubber transition analysis of an amorphous food powder.

Table 1 T_g-r of whey, honey, and apple juice powders as a function of moisture content

Sample	Moisture (% dry basis)	Tg-r (°C)
Whey Powder	2.77 ± 0.12	69.4 ± 0.8
	5.84 ± 1.22	61.8 ± 0.1
	9.03 ± 1.41	37.7 ± 0.4
Honey Powder	2.68 ± 0.36	58.5 ± 0.6
Apple Juice Powder	3.38 ± 0.18	55.3 ± 1.4

Table 2 Transition and sticky temperatures (°C) of honey and apple juice powders measured by TMCT, DSC, and TMA techniques

			Tg (DSC)
Sample	Moisture (% db)	Tg-r (TMCT)	Tg onset	Tg midpoint	Tg endset	Tonset (TMA)
Honey Powder	2.68 ± 0.36	58.5 ± 1	53.9 ± 1	54.8 ± 1	57.5 ± 1	88.7 ± 2
Apple Juice Powder	3.38 ± 0.18	55.3 ± 1	50.3 ± 2	56.4 ± 1	60.7 ± 0.5	81.7 ± 4

For comparison, glass transition analysis using standard DSC and TMA techniques was also carried out. Fig. 3 compares the results obtained by the new TMCT technique with the standard TMA and DSC techniques for whey powder. The T_g-r values were found 14.5 ± 6°C above the T_g. Even though T_g-r values were expected to be closer to T_g, the results were reasonable, since the presence of large amount of proteins might have masked the small mechanical changes due to glass transition of amorphous lactose at the initial stage. For all moisture levels, T_g-r values were lower than T_onset measured by TMA. This indicates that the TMCT technique is more sensitive to detect the state change. The higher TMA results suggested the temperature lag between the oven temperature and the sample temperature, which is the common nature of the DMTA instrument. Moreover, this could have also been the effect of porosity on the transition measurement as found by Roos et al.[6] Comparing between the transition temperatures of porous and tablet extruded samples, Roos and coworkers found that the T_g for tablet sample was much lower than the porous sample,[6] and it was believed to be as the result of an inversed relationship between porosity and thermal conductivity of the sample.[6,19] In this study, the porosity of the sample bed under TMCT would be much less than that of the sample under TMA test due to the fact that the compression force for TMCT was 3 times higher. Therefore, the results obtained from the TMCT well agreed with the porosity concept, which stated that the more porous material would take longer time to reach a specified temperature and hence resulting in a detection of higher glass transition temperature.[6,19]

Figure 3. Comparison of transition temperatures measured by TMCT, DMTA, and DSC for whey powder.

Result comparison for honey and apple juice powders is presented in Table 2. It was found that T_g-r was very close to T_g measured by DSC. For both samples, mechanical changes at the surface of powder bed in contact with the temperature controlled sample holder expressed in terms of T_g-r was determined at temperature closer to T_g midpoint. The T_g-r values for honey and apple juice powders having moisture contents of 2.7 ± 0.4 and 3.4 ± 0.2% dry basis were 58.5 ± 1 and 55.3 ± 1°C, respectively. Because they contained sugars (sucrose, glucose, and fructose), which have much lower T_g values (62, 31, and 5°C, respectively) than lactose (T_g = 101°C) in the dairy powder,[20] they were more sticky and difficult to produce by spray drying, even when maltodextrin was incorporated. For both samples, the results obtained using the TMCT and the standard techniques were in agreement. This suggested that the novel TMCT technique was successfully applied for the measurement of stickiness behavior in food powder, and they were valid for this purpose.

Application of TMCT for Melting Point Analysis

The melting points for glucose monohydrate and sucrose crystals measured by the TMCT, DSC, and DMTA/TMA techniques are presented in Table 3.

Table 3 Melting points of glucose and sucrose crystals measured by different techniques

			DSC
Sample	TMCT Tm (°C)	TMA Tonset (°C)	Tm onset (°C)	Tm mid-point (°C)	Tm endset (°C)
Glucose monohydrate	156.6 ± 1.0	155.2 ± 2.0	153.3 ± 2.0	157.7 ± 1.0	160.0 ± 0.5
Sucrose	194.6 ± 1.0	196.4 ± 1.0	194.3 ± 2.0	200.4 ± 2.0	203.8 ± 2.0

The results among different techniques were comparable, as there was no influence from any other component when the crystals were tested. Moreover, it was found that the probe displacement-temperature data may not need thermal expansion correction using the standard maltodextrin[1 for melting point analysis, since the melting occurred over a narrower temperature range and exhibited a more rapid change in mechanical property as compared to glass transition. There was only a single sharp change in mechanical behavior during the melting point analysis as the crystals were pure components. The results also agree with the reported temperature ranges.[20–21 This study shows that the new TMCT technique was capable of measuring physical transformations that involve changes in mechanical properties.

CONCLUSIONS

The TMCT technique was applied for measurement of the glass-rubber transition temperature of some amorphous powders, which exhibit stickiness problems during the processing, handling, and transporting operations. Comparisons were also made on the glass transition analysis results obtained from TMCT, DSC, and DMTA tests. It was found that the sensitivity of TMCT in measuring mechanical property changes during glass transition was higher than the standard measurement of heat flow in DSC technique. Large differences found between T_g-r measured by TMCT and T_onset measured by TMA were mainly due to temperature lagging nature of the latter.

ACKNOWLEDGMENT

The sponsorship from Chiangmai University, Thailand, and The University of Queensland is gratefully acknowledged.

Notes

12. Boonyai, P. Development of Instrumental Techniques for Measurement of Stickiness of Solid Particulate Food Materials. Ph.D. Thesis, The University of Queensland, Brisbane, Australia 2005; 239 pp